Underwater acoustic (UWA) communication faces stringent constraints on bandwidth, power consumption, and transmission reliability due to the challenging acoustic propagation environment. This project aims to develop a novel AI framework for source compression and semantic communication in UWA communication networks. Specifically, high-level semantic representations—such as geographic location, environment mapping, and mission-relevant features—shall be extracted and can reconstruct the raw data (image or video). This approach aims to significantly reduce transmission load, improve robustness to channel noise, and enable multi-task semantic communication for essential underwater operations.
The proposed solutions will be implemented using Python and software-defined radio platforms. The expected outcomes include a validated AI-driven compression framework for realistic underwater acoustic channels and a demonstration of semantic communication efficiency in bandwidth-limited scenarios. Multiple field tests will be conducted, with the collected data serving as the foundation for further optimization of the AI framework.
School
Electrical Engineering and Telecommunications
Research Area
Acoustic communication | Machine learning
Suitable for recognition of Work Integrated Learning (industrial training)?
No
- Research environment
- Expected outcomes
- Supervisory team
- Reference material/links
Python and software-defined radio platforms, underwater field test
AI-driven compression framework for realistic underwater acoustic channels, a Python demo of semantic communication efficiency, software-defined radio test-bed, report/publication writing
- Goodfellow I J, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets[J]. Advances in neural information processing systems, 2014, 27.
- Zhou S, Wang Z. OFDM for underwater acoustic communications[M]. John Wiley & Sons, 2014.